Update 2d_appimage.sh

[johnnynunez]

more robust

[johnnynunez]

more robust

I've can build on raspberry pi and jetson agx orin :)

[alemuntoni]

Hi @johnnynunez , thank you for the PR!
The idea is to build meshlab (and then pymeshlab) on linux arm, right?
Right now I don't have any arm machine on which I could make tests, but it would be very useful!
I saw somewhere that github actions planned linux arm free runners, therefore arm builds could be added in our releases.
But right now it is difficult for me to patch and test meshlab on arm :(

[johnnynunez]

Hi @johnnynunez , thank you for the PR! The idea is to build meshlab (and then pymeshlab) on linux arm, right? Right now I don't have any arm machine on which I could make tests, but it would be very useful! I saw somewhere that github actions planned linux arm free runners, therefore arm builds could be added in our releases. But right now it is difficult for me to patch and test meshlab on arm :(

I can create it with arm!
I've using jetson-containers from Jetson NVIDIA.
I can build meshlab and pymeshlab with support GPU!
CONTAINER: https://github.com/dusty-nv/jetson-containers/tree/dev/packages/nerf/meshlab
CONTAINER: https://github.com/dusty-nv/jetson-containers/tree/dev/packages/nerf/pymeshlab
APP: https://forums.developer.nvidia.com/t/meshlab-2023-12/305249
WHL: http://jetson.webredirect.org/jp6/cu122/pymeshlab/2023.12.post1

[alemuntoni]

Great! Does it mean that the only thing to do was to use the arm linuxdeploy version and then meshlab built and run out of the box?

[johnnynunez]

Great! Does it mean that the only thing to do was to use the arm linuxdeploy version and then meshlab built and run out of the box?

yes, linuxdeploy(that I build manually because not exists for arm either) and appImage for arm64 for create meshlab app

wget https://github.com/AppImage/appimagetool/releases/download/continuous/appimagetool-aarch64.AppImage \
    && chmod +x appimagetool-aarch64.AppImage \
    && mv appimagetool-aarch64.AppImage /usr/local/bin/appimagetool

so if you want to use my files of linuxdeploy and create two folder x86 and arm64 feel free.
I mean is possible to detect the architecture and execute appimage and linuxdeploy for this architecture

[alemuntoni]

yes, linuxdeploy(that I build manually because not exists for arm either) and appImage for arm64 for create meshlab app

Did you check the linuxdeploy github repo? They have builds also for arm64: https://github.com/linuxdeploy/linuxdeploy/releases

Anyway, the Dockerfile in the .devcontainer directory is not related with the github actions runner.
Why do you say that appimagetool must be in the PATH? Only the plugins for linuxdeploy should be required, and it is fine if they are in the same directory of linuxdeploy

[johnnynunez]

wget https://github.com/AppImage/appimagetool/releases/download/continuous/appimagetool-aarch64.AppImage
&& chmod +x appimagetool-aarch64.AppImage
&& mv appimagetool-aarch64.AppImage /usr/local/bin/appimagetool

yes, but not all plugins are build

[alemuntoni]

yes, but not all plugins are build

I see. linuxdeploy-plugin-qt does ship for arm, but linuxdeploy-plugin-appimage does not at the moment.

[alemuntoni]

Anyway, I still don't see why appimagetool should be available in the deploy script.
Just the linuxdeploy-plugin-appimage should be required.

[johnnynunez]

@alemuntoni ready for merge

[johnnynunez]

I see. linuxdeploy-plugin-qt does ship for arm, but linuxdeploy-plugin-appimage does not at the moment.

That's right, and I have compiled the repositories for aarch64, for the jetson agx orin and future jetson thor.

[johnnynunez]

@alemuntoni Now it is possible to add CI ARM64 directly on github actions:
https://github.blog/news-insights/product-news/arm64-on-github-actions-powering-faster-more-efficient-build-systems/